Mainframe may be an unknown name to gamers, but the founders are well respected veterans in the industry. Composed of two offices, one in Reykjavík, Iceland and another in Helsinki, Finland, Mainframe was founded by former CCP Games and Remedy Games developers with the highly ambitious goal to make an open-world social sandbox MMO game that would run entirely in the cloud.

　　After an initial seed round of two million euro announced in October, Mainframe has now managed to secure another €7.6 million in Series A funding led by Andreessen Horowitz (a16z) with additional investment from Riot Games.

　　Following that news, we were able to schedule a long and interesting chat with CEO Thor Gunnarsson (previously VP of Business and Development at CCP Games), who explained why the folks at Mainframe are convinced that the industry is headed towards cloud native games in the next few years.

　　Let's start with the elephant in the room, COVID-19. With the fact that you have had two offices since the foundation of Mainframe, one in Helsinki and one in Iceland, perhaps you are used to remote collaboration already.

　　That's right. I suppose we were fortunate in that sense that from the founding of Mainframe in April of last year, we already began working as two parallel teams, with our founders in Helsinki and the founding team here in Iceland. So we've been practicing and have pretty good remote work methods already between the two offices. Now, of course, it's all individuals working from home but the first couple of days of this seem to be working reasonably well.

　　What do you think the impact of COVID-19 will be on the games industry as a whole?

　　It's a tough question that's of course top of mind for everyone in the games industry. I think on the one hand we know that games are a sort of a countercyclical industry in times of recession or economic hardship. People tend to spend more time with games, on balance it is a less expensive form of entertainment compared to other ways of entertaining ourselves. We certainly saw that trend when I was at CCP Games in 2008 during the financial crash. There, we saw increasing usage and increasing player numbers throughout basically the downturn then. Today, of course, it's a little more unpredictable in the sense that we have not in modern times gone through quite as dramatic a public health crisis as we are now. So it's a little bit difficult to predict and of course, challenging for any industry to accurately predict one way or another. But we hear from our peers in the industry that game usage in the short term is certainly is up.

　　Yeah, because people don't have much to do besides Netflix and gaming.

　　That's right. And unlike Netflix, which, of course, is quite a passive way of consuming entertainment, games and online games are inherently social activities.

　　Games are where we will find and maintain our friendships over the coming months. So in that sense the games industry is providing not just an entertainment option, but really a social platform for people to connect and stay in touch with each other. And I think that aspect of online gaming is obviously tremendously valuable under the challenging circumstances.

　　Particularly MMOs, because those are basically virtual worlds in their own right.

　　Those are perhaps unique in some ways, relative to the broader online category in that the social network that forms within successful MMOs leads to kind of longer standing social connections and friendships and you know, sometimes obviously, enemies. Over time, a successful MMO that has a very strong social layer really is sort of the genre of online game that perhaps has the highest kind of social utility for its players. Reflecting on this from our time at CCP, when we were working on EVE Online, we saw tremendously valuable personal connections and friendships and you know, on many occasions even weddings that happened as a result of people meeting in EVE Online. So that element, I think, is in this kind of challenging times even more important for people to find their social place. And perhaps find a place that is not full of a newsfeed of Coronavirus news as we're seeing on the conventional social networks and social media. So MMOs are potentially a form of escapism, but also perhaps a respite from the real world on occasion.

　　Right. Let's talk about Mainframe. I know you have a relatively small team, but are you looking to grow significantly in the coming years?

　　We are looking to grow. The current team size is 20 developers, equally split between Helsinki and Reykjavik. With the news of the investment from Andreessen Horowitz and Riot Games, we are certainly looking to grow our team in both locations. I would say, broadly speaking, we're looking to add about 30 people this year to the team. We're not looking to build a studio of hundreds of developers. We think that there is a kind of a size of team that particularly experienced game developers can kind of strive to actually do pretty significant development with a headcount well under 100 people. So we're not looking at creating a very large team for Mainframe, but we're really kind of leaning into the experience of the founding team and hoping to attract talented and experienced game developers and designers from across the Nordics and obviously internationally for the game that we're making.

　　Since you are discussing a relatively agile team as your goal for Mainframe, I'm wondering if you're going to focus on more sandbox, emergent gameplay, rather than theme park content, which is a lot more expensive in terms of manpower.

　　Yeah, you kind of hit the nail on the head with respect to the relative difference between a very content heavy, themepark style MMO and the other kind of design like a sandbox MMO game that is designed for emergence. Our pedigree with the team here in Iceland, this is a lot of the senior team that actually created EVE Online back in the day. Guys like our Game Director, he was the co-founder and creative director of CCP, and our CTO was the lead game designer of EVE back in the day. So we have quite a few people that certainly have that pedigree, and kind of a deep understanding of what's needed to create the framework that allows a sandbox world to begin to emerge. And to kind of design a game that is not overly prescriptive or overly designed, but more open to player action and player behavior. And our job if we execute it well is to create a world that is reactive to the play styles and the ways in which our players actually inhabit our world. So our job in that sense is to kind of listen, learn and be, as you say, very agile in our development, reacting to kind of how our player community evolves over time.

　　Do you also plan to have one big single shard server architecture like in EVE Online?

　　It's a good question. We're not kind of giving that amount of detail about the game at this stage. But, you know, at Mainframe we certainly have a keen appreciation of both the positives and challenges of doing a single shard architecture. And it's something that frankly, we're evaluating and we'll have more information on as we're able to say more about the game in the future.

　　And, of course, the big thing about your game is that it is going to be cloud native, right?

　　That's right. We founded Mainframe really with a keen belief that cloud native games are really where our industry is headed. This is not a development that will happen literally over the course of a year. But over the coming years, we're going to see quite a dramatic change in both how games are made, and how games are experienced and how players engage with their favorite games. And we think that particularly for a team that has high ambitions to create a sort of an MMO that reimagines, in some ways, how MMOs and the social aspects of MMOs work. We think that the ability to create a world that can be accessed by our players from any device and any screen is incredibly compelling. It allows us to create a world that can span from very short, casual moments where you may check in on the world from your phone to more in-depth based sessions when you're at home in front of your PC or your console. When we began to understand how many of the large platform companies are moving with streaming services with a focus on cloud gaming, we saw an incredible opportunity to form a team and kind of advance some game ideas that we've had for many years but simply weren't possible in kind of the previous client server architecture that we used to work within.

　　Let's say you have a high-end PC. In this case, is the game still going to act just like a client that accesses streaming, or are you going to take advantage of the local hardware in some ways?

　　It's a good question where particularly for players that have a high-end PC, there are certainly pros and advantages to being able to take advantage of local rendering resources. And we see some of the technology companies that are active in this area are taking this sort of hybrid approach. From our perspective, if we think about some of the innate values of streaming, it's not just about accessing the game remotely or streaming that to your device, it also has some other benefits to both gamers and to the creators of those games. For example, when the client is running in the cloud, it's very difficult to hack that client because you don't have access to it, what you're receiving essentially is a video frame. So there's no spying the map, wallhacks become almost impossible. A tremendous amount of time, particularly in the MMO space, goes into anti hacking and anti cheat measures. And we see that across all online games. So with a game that lives entirely in the cloud, the benefit is quite high, you know, both to us in terms of reduction in complexity and kind of anti-cheat measures but also for our players as it means that maps can be genuinely unknown. It means that the world that is constantly changing and evolving, of course, we are creating a persistent world in the cloud. It means that the way in which players experience virtual worlds is going to change dramatically because it's not possible to simply hack the client and understand all elements of the map.

　　There would be no data mining either, right?

　　Very, very little data mining. I should be clear to add 'Never say never'. Hackers are very clever and they will possibly find ways to do things like image recognition on the video stream to try and understand what's happening in the world, etc. But to us, you know, as long standing developers in the MMO space, we understand how challenging this is and how many constraints it places upon game design. And by removing just that element, it just dramatically changes the kind of game we actually make. So that's kind of one example of what to us feels like a pretty compelling shift enabled by cloud gaming. But there are certainly other aspects of cloud networks, the ability to access the world in a friction-free way, that is incredibly compelling to us. It means that as the game developers, we can create worlds where people can access those worlds simply as passive viewers. So whether you're on Twitch or YouTube or Mixer, you begin to understand what the world is about, potentially then, you know, jump into the world with a click from your friend or from an influencer that is offering you access to the game.

　　And now you're suddenly right in the action, you're possibly in the middle of a raid party where you're actually making a meaningful contribution to an encounter that is happening live and being watched by hundreds of thousands of people. And moreover, when you are creating a game that is so inherently social as a sandbox MMO, when you as a player of the game are thinking about how you recruit your friends to your cause in the game, to be able to send that link to your friend over social media or email or SMS and have them come directly into the game and be sort of introduced to the game by you personally and your clan is incredibly powerful. It dramatically changes how people have their first experience with online games. And ultimately it leads to what we at Mainframe believe will be a much stronger kind of explosion of social activity and ultimately a social network forming within that game.

　　In your first press release, there was a statement from yourself, actually, where you said that the cloud has the potential to allow for bigger and better worlds, with more lively AI and improved physics and simulation. That's something that I'm also very interested in personally as a gamer because I think usually in MMOs these are aspects that are behind the regular triple-A gaming experience. Usually, AI and physics are not that great in MMOs. Can you talk about that?

　　Yeah, I can expand a little bit on that. I think as an industry and as game developers we're all getting to grips with both the technical and creative possibilities that the cloud offers. But some obvious opportunities are, for example, the increased storage capacity that you can have if you're running the game and the simulation of the game in one central location. Being able to rely upon much faster and greater storage capacity is a dramatic change in how we can create a world. So let's take a hypothetical example, if you're a player character in the game, over the course of your journey in the game you are potentially aging, you're potentially taking damage or even receiving scars from combat. Those scars could be personal to your character and can be seen by all other players, which is simply not possible in the current client server architecture where the game always has to rely upon the available memory and bandwidth of the lowest common denominator client connecting to it. That's just one simple example.

　　But the ability also to create larger worlds, larger open world environments where the sort of transition from one area of the world to the other can be made much more seamless in the cloud than if we were working through, again, a traditional client server environment where we have to load that data and ultimately do that transition not just for the individual player but for everyone else who's in the world with them. Distributed physics simulation, of course, is a massive problem in online games. When you're able to run the simulation in a central location and then distribute the results of that simulation to all of the streaming clients that are connected, we think we can also do much more with physics than certainly we've been able to accomplish today. So there are a lot of exciting ways to kind of leverage that going forward.

　　Actually, that's something that Google also discussed as a possibility for their Stadia platform. They called it elastic rendering, where developers can use two servers instead of one to render a single instance of a game. This is one of the greatest positives of cloud gaming, but as you know, streaming is also still in its infancy in a lot of ways. For example, Google is having quite a few issues since the launch of Stadia, mostly related to things like latency, for example, or even the graphical quality, which is not exactly as promised by Google. What do you think about that, is it just going to be an issue still for some time?

　　The way in which I think about this, personally at least, is we're at just the kind of year zero of the cloud gaming shift. It's very similar in my mind to what we all experienced when we were beginning to play online games back in the day, when most players were on a dial-up modem playing kind of early in software games. The early kind of people that were beginning to play MMOs were connecting over wildly varying devices. Over some years, broadband became much more prevalent and the experience became much better for everyone. But I think it is certainly true that cloud gaming and particularly streaming, whether it's over a fixed line broadband or over 5G will take some time to reach all of the addressable market of potential players. So there will be a period of time where connectivity, bandwidth and so on will be obviously variable. But it's not tremendously different from what we saw with the advent of online gaming 20 years ago when we were seeing that slow kind of progression towards always-on broadband and so on.

　　And I think we're sort of in that transition phase with cloud gaming today, so again, this won't be a cloud revolution in the year 2020. But, you know, two, three years from now, we certainly think that there is going to be quite a dramatic shift. And one of the ways in which I think that will happen is really with the rollout of 5G networks across the world. We've seen in some of the technical tests that we've had a chance to undertake in Finland latency characteristics that are just incredibly impressive, with very low ping times to the edge of the network. And when you look at what companies like Amazon with their AWS initiative are doing to bring streaming and cloud gaming to the edge of the mobile operator network, that is dramatically useful in terms of gaming experience. And that, of course, is one of the key enablers for actually bringing cloud gaming to mobile phones. We know that in certain parts of the world 4G LTE is pretty awesome. But you go to other corners of the world, and it certainly is overloaded, the capacity is not good and latency is very variable. It sort of fluctuates, like a sine curve. With what we've seen on certain early 5G networks, that kind of problem is dramatically diminished. And so we believe that we can actually bring a triple-A, visually immersive, compelling experience to a mobile device because of 5G and that rollout that will be happening over the coming years.

　　Can you divulge which engine you're using for your first game at Mainframe?

　　Yes, we are an Unreal studio all the way. Phenomenal platform. We've been working with it in past companies over a number of years, and we absolutely love Unreal Engine.

　　Since your game is going to be future-proofed in a way, are you going to use technologies like ray tracing?

　　I hope so. You know, ray tracing in the cloud is just incredibly impressive. Looking at what companies like NVIDIA are already doing by upgrading GeForce Now to support RTX in the cloud, is just incredibly compelling. Because, again, if we can have a single runtime target in the cloud, let's say for sake of argument that's equivalent to a 2080 high-end Nvidia GPU as our baseline of performance, we can deliver that visual experience to any device. That is just mind blowing as an opportunity for a game developer, to be able to target one single runtime rather than doing a multiplatform porting strategy where your graphics and your kind of experience is only as good as the lowest common denominator platform that you're targeting. In that sense, actually from a game developer perspective, it not just gives us creative and technical opportunities for our game but also reduces our cost and complexity of development. And that's kind of another major bonus in terms of moving to this cloud native posture.

　　One additional question from a general developer point of view. I'm not sure if you noticed the recent Xbox Series X reveal. They discussed that machine learning (through DirectML) is going to be a big part of the console and I believe the same will go for Windows PC games. Are you evaluating this opportunity at Mainframe?

　　We're looking at a number of different ways to take advantage of AI. I don't have a specific take on Microsoft's news on that, but the broad category of machine learning and generally speaking AI is something that we're looking at both in terms of how we create our worlds, and how we simulate and kind of bring life to those worlds.

　　That's also something that can improve development time, right?

　　Yeah, absolutely. We're very much leaning on the latest technologies and engines that are available to us to create the world and then, of course, to make that world alive. So procedural world generation is something that we're looking at very closely and already using in kind of early prototyping of the game. It's incredibly promising. There's so much technological advancement and particularly, again, when you can say 'Look, I'm targeting one single runtime', now I can make choices in my game design and in my technology stack that would have been very difficult if I were doing, again, a conventional multi-platform client server game. So that just is hugely exciting.

　　This is the second round of financing for Mainframe. Do you have plans to do any further rounds?

　　We're in very good shape based on this new round that we have announced with Andreessen Horowitz and Riot Games. We will doubtless be adding more weight behind the company over time, but for the time being, this series is basically taking us forward to some quite meaningful milestones in terms of kind of first player access to our game and so on. Don't have more news on the overall kind of funding and how that will play out over the next coming years. But we're certainly incredibly happy to bring a16z and Riot to the company and kind of the value that we know they will add to Mainframe is just tremendous. To be able to spend time with the team at Riot discussing game design, live operations and kind of their incredible expertise in running the largest PC gaming in the world is priceless for us as a company.

　　When do you think we can expect to hear the first details on the game that you are working on?

　　I wish I could say more about that at this point. We're going to have more news about the game, obviously, in due course. But we're a little bit away from that reveal at the moment.

　　All right. Well, I'm personally looking forward to it. Personally, I could never quite get into EVE Online because it's quite the abstract gameplay experience. But I love the concept behind it and the sandbox elements, so if you can put that into a more accessible, action-oriented experience, then I'm all for it.

　　It's exactly what we're setting out to achieve, that really is the DNA of the studio that we worked so hard to put together is really combining some of the expertise that we have learned from the CCP guys with just incredible talent in Helsinki from the triple-A PC and console side, with a number of folks from Remedy joining company. And then the kind of expertise in user interface and user experience coming from some of our team that were previously doing incredibly accessible and social mobile games. And we're hoping to kind of bring that all together into an experience that certainly is less abstract than EVE Online was.

　　Splendid. Thank you for your time.